> Date: Tue, 17 Aug 1999 10:43:56 -0700
> From: Cindy Ballreich <[EMAIL PROTECTED]>
> I thought it would be interesting to try out the new compression utilities
> by feeding the same GeometryInfo object into a CompressionStream object
> and generating a CompressedGeometry object from the data. When I displayed
> the resulting object, I found that the geometry had been scaled down by
> 0.002 [...]
This part is expected behavior. The geometry has been quantized and scaled
to a normalized cube with the open endpoints (-1..1) as part of the
compression process. The maximum range of the resulting data will be one
quantum away from unity at both ends of the range; for example, at a
position quantization of 6 bits, an object would be normalized so that its
most negative dimension is at (-1.0 + 1/64) and the most positive is at
(1.0 - 1/64).
This wastes one quantum of resolution, since 2's complement integer
arithmetic does allow us in fact to express -1.0, but not +1.0. If we were
to use this extra quantum the scale and offset wouldn't be symmetrical about
the origin, so we decided to keep it symmetrical.
Unfortunately, the compression utilities do not provide a method to return
the exact bounds to the application, and they probably should. For now you
can compute the bounds according to the above formula, and we will see how
best to overcome this shortfall in the API in a future release, hopefully
Java 3D 1.2.
> [...] there were some nasty artifacts (spots and black lines that vary
> depending on the data) in the vertex colors. Does anyone have any thoughts
> about what I might be doing wrong? (The same data displays correctly as a
> TriangleStripArray.) Here's a bit of code...
This part is not expected behavior and could indicate a bug in that we have
not yet encountered. Your code snippet looks fine, but I'd greatly
appreciate it if you could email me your data and the particulars of your
setup, and perhaps a capture of the image that you're getting. I'll
generate a bug report and work on a fix once I can reproduce the problem.
Thanks for sending us a report of your experiences with compressed geometry.
-- Mark Hood
[EMAIL PROTECTED]
Java 3D development team
===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff JAVA3D-INTEREST". For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".
